# Empathy Is Steerable but Multi-Axial: Mechanism Geometry and Persona Effects in LLMs

arXiv: https://arxiv.org/abs/2609.15654 · HTML: https://mlsanity.com/paper/2609.15654/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Activation steering methods, one of 13 directions in Mechanistic interpretability.
- Builds on 20 of its 30 references that are in the map.
- First posted on arXiv 14 Sep 2026.

## Placed in

- Mechanistic interpretability / [Activation steering methods](https://mlsanity.com/mechinterp/activation-steering-methods/)

## Builds on (most cited first, 20 of 20)

- [Representation Engineering: A Top-Down Approach to AI Transparency](https://mlsanity.com/paper/2310.01405/) (2023)
- [Inference-Time Intervention: Eliciting Truthful Answers from a Language Model](https://mlsanity.com/paper/2306.03341/) (2023)
- [Steering Llama 2 via Contrastive Activation Addition](https://mlsanity.com/paper/2312.06681/) (2023)
- [Refusal in Language Models Is Mediated by a Single Direction](https://mlsanity.com/paper/2406.11717/) (2024)
- [Steering Language Models With Activation Engineering](https://mlsanity.com/paper/2308.10248/) (2023)
- [The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasets](https://mlsanity.com/paper/2310.06824/) (2023)
- [Towards Emotional Support Dialog Systems](https://mlsanity.com/paper/2106.01144/) (2021)
- [A Computational Approach to Understanding Empathy Expressed in Text-Based Mental Health Support](https://mlsanity.com/paper/2009.08441/) (2020)
- [CEM: Commonsense-aware Empathetic Response Generation](https://mlsanity.com/paper/2109.05739/) (2021)
- [Evaluating Large Language Model Biases in Persona-Steered Generation](https://mlsanity.com/paper/2405.20253/) (2024)
- [On the steerability of large language models toward data-driven personas](https://mlsanity.com/paper/2311.04978/) (2023)
- [FairSteer: Inference Time Debiasing for LLMs with Dynamic Activation Steering](https://mlsanity.com/paper/2504.14492/) (2025)
- [Large Language Models Produce Responses Perceived to be Empathic](https://mlsanity.com/paper/2403.18148/) (2024)
- [When Large Language Models are Reliable for Judging Empathic Communication](https://mlsanity.com/paper/2506.10150/) (2025)
- [Personality Vector: Modulating Personality of Large Language Models by Model Merging](https://mlsanity.com/paper/2509.19727/) (2025)
- [Valence-Arousal Subspace in LLMs: Circular Emotion Geometry and Multi-Behavioral Control](https://mlsanity.com/paper/2604.03147/) (2026)
- [HEART: A Unified Benchmark for Assessing Humans and LLMs in Emotional Support Dialogue](https://mlsanity.com/paper/2601.19922/) (2026)
- [PERSONA: Dynamic and Compositional Inference-Time Personality Control via Activation Vector Algebra](https://mlsanity.com/paper/2602.15669/) (2026)
- [Are LLMs Empathetic to All? Investigating the Influence of Multi-Demographic Personas on a Model's Empathy](https://mlsanity.com/paper/2510.10328/) (2025)
- [Language Models Predict Empathy Gaps Between Social In-groups and Out-groups](https://mlsanity.com/paper/2503.01030/) (2025)

## Abstract (arXiv)

Activation steering has been used to control traits such as honesty, refusal, and sycophancy, yet supportive empathy is evaluated along multiple dimensions that need not correspond to independently controllable activation directions. Using the EPITOME framework, which decomposes supportive empathy into Emotional Reactions, Interpretations, and Explorations, we study three instruction-tuned LLMs and ask whether candidate directions derived from these labels produce distinguishable intervention effects or instead share structure, and how persona prompts interact with those directions. We find that contrastive activation addition yields a stable middle-layer intervention that consistently shifts the EPITOME proxy scores across models, moving empathy analysis beyond response-level scoring. However, the recovered directions are only partially separable: steering one direction induces off-target shifts, and hand-crafted prompting shifts the empathy profile rather than isolating a single dimension. Persona prompts substantially change EPITOME scores, but a paired activation-shift decomposition shows that the recovered subspace captures only approximately 3 percent of persona-induced squared activation-shift magnitude at layer 15. Under this EPITOME-based definition, expressed empathy is steerable but multi-axial, and controlling persona-conditioned empathy requires targeting structure beyond individual mechanism directions.